Ongoing Research on Herding Agents for In Situ Burning in Arctic Waters: Laboratory and Test Tank Studies on Windows-of-Opportunity

This paper provides an overview and summarizes the findings from one part of a research program on increasing the knowledge of chemical herding agents: their fate; their environmental effects; and, the windows-of-opportunity to expand the operational utility of in situ burning (ISB) in cold open water and loose drift ice conditions. ISB of accidentally spilled oil is a response technology with great potential, particularly in ice covered waters. One of the key parameters of effective ISB is the thickness of the oil slicks: a thick slick is required for ignition and efficient burning (Buist et al. 2013). Several methods exist to keep an oil slick at the required thickness for burning, e.g. fire resistant booms and ice edges. In loose drift ice conditions (approximately 10–60 % ice coverage) oil spills can rapidly spread to become too thin to ignite; however, the ice concentration is too high for fire resistant booms to be deployed and operated efficiently. Herding agents and their ability to contract and thicken oil spills in loose pack ice for in situ burning has been described most recently in a research summary report (SL Ross and DCE 2015) where the authors summarize their history in oil spill response and recap the last 12 years of research.

Herders have the ability to spread rapidly over a water surface into a monomolecular layer as a result of their high spreading coefficients, or spreading pressure. The best herding agents have spreading pressures in the mid-40 mN/m range, whereas most crude oils have spreading pressures in the 10–20 mN/m. A monomolecular layer of herding surfactants reduces the surface tension of water from 70 mN/m to 20–30 mN/m. When the herders contact the edge of a floating oil slick they change the balance of the interfacial forces acting on the slick thereby allowing the interfacial tensions to contract the oil. Only small quantities of herder are needed, e.g. 150 μl/m² in contained areas and 15 L/km at sea, are the recommended application dosages to clear thin films of oil from large areas of water surface. Recent research (Potter et al. 2017 in these proceedings) has examined herder application from helicopters.

The research project involved several separate studies in different locations:

• Laboratory studies on the physical fate of herders during ISB at DTU (Technical University of Denmark) and DCE (Danish Centre for Environment and Energy, Aarhus University).

• Environmental effects studies on: toxicity testing and bioaccumulation of herders with Arctic copepods; the potential for biodegradation in Arctic conditions; and, the impacts on Arctic seabirds (Thick-Billed Murre and Common Eider) at DCE, and in Greenland.

• Chemical analysis of the smoke generated during ISB of oil slicks confined with herders at the SL Ross laboratory in Canada.

• Windows of opportunity for effective use of two herders on a range of crude oils (ANS, Endicott, Grane and Terra Nova) at a small-scale (SL Ross lab in Canada) and larger-scale (USACE-CRREL - Cold Regions Research and Engineering Laboratory, Hanover, NH).

This paper focuses on the Windows of Opportunity experiments. The results of the fate and environmental effects research are presented in a separate paper in these proceedings (Fritt-Rasmussen et al. 2017). More details on the complete study results can be found in the technical report (Buist et al. in prep.).

The two herders used in this project have been placed on the U.S. EPA National Contingency Plan (NCP) Product Schedule list and are commercially available. ThickSlick 6535 (TS6535) is a blend of 65 volume % sodium monolaurate (Span 20), the surfactant, and 35% 2-ethyl butanol, the solvent. The active ingredient of TS6535 is used as a food additive, in household cleaners as well as in cosmetics, fine fragrances and other toiletries. OP-40 is a proprietary polydimethylsiloxane copolymer. Surfactants of the type used in OP-40 are used in household and automotive care products as well as in hair conditioners and skin care products.

Four crude oils were selected for these tests:

• Alaska North Slope pipeline crude (sample obtained from a refinery in California in 2013)

• Endicott crude from the North Slope of Alaska (sample obtained in 2004)

• Grane crude from a platform off Norway sourced in 2014.

• Terra Nova crude from a 2002 tanker load taken from a production FPSO off Canada.

The small-scale herding experiments were conducted at the SL Ross lab (Figure 1) in:

• 1-m² metal pans lined with a rinsed clean plastic film;

• Metal pans on a rocking shaker used to simulate wave action called the Dynamic Film Performance (DFP) test;

• A 10-m² pool created on the floor of the lab using lumber and clean plastic film; and,

• An 11 m × 1.1 m × 1.1 m refrigerated wind/wave tank.

The crude oils for the experiments were artificially evaporated by sparging compressed air into pre-weighed 20-L cans of the oils. The target evaporative losses were based on earlier spill-related property studies of the four crudes.

The 1-m² pan tests were done with 0°C, 35 ‰ salt water and included as variables:

• Two herding agents (TS6535 and OP-40);

• Four crude oil types (Alaska North Slope, Endicott, Grane and Terra Nova);

• Three extents of evaporation for each crude (fresh and two degrees of evaporation representing a few hours and a few days at sea);

• Two emulsion water contents (20% and 50%) for each crude that could form emulsion; and,

• Slush ice in selected tests.

The Dynamic Film Performance (DFP) experiments were done with 0°C, 35 ‰ salt water in a small environmental chamber to maintain temperatures at 0°C. The test matrix included:

• Two herding agents (TS6535 and OP-40)

• Three crude oil types (Alaska North Slope, Grane and Terra Nova);

• Three extents of evaporation for each crude (fresh and two degrees of evaporation); and,

• Two emulsion water contents for each crude capable of forming emulsions.

The 10-m² experiments were performed on the concrete lab floor in a rectangular, wooden frame (3.05 m × 2.95 m × 9 cm deep) lined with a new, rinsed plastic sheet of 1-mil plastic film to ensure a clean, uncontaminated surface. The 10 m² experiments involved ANS and Grane crude (fresh and evaporated), with some tests incorporating slush ice while others used emulsions.

The results are presented here for one of the oils tested, ANS crude, as it was involved in all four test apparatus. Generally the tests with the other three oils produced similar results. Where appreciable differences with other oil types were observed, these are noted. Full results are given in the project final report (Buist et al. 2017 in prep.).

Figure 2 shows the results of the experiments with fresh and evaporated ANS crude on cold salt water. The experiments with the OP-40 herder are shown in shades of red on the left of the chart, and those with ThickSlick 6535 (TS6535) are shown on the right of the chart in shades of blue. Evaporated ANS crude would not form a sufficiently stable emulsion for these experiments.

The initial spreading of the fresh ANS crude was quite typical of what has been observed in other experiments with this crude oil with equilibrium thicknesses of the crude of 0.6 to 0.7 mm (duplicate runs) in the 1-m² pan. The OP-40 herder was more effective than the TS6535 herder with fresh ANS crude, achieving a herded thickness of 4.4 to 4.6 mm one minute after application. This thickness declined over the one-hour test period to an average of 3.1 mm (2.6 to 3.7 mm for the duplicate runs). In comparison, the TS6535 herder achieved a herded thickness of 1.9 mm that slowly increased to 2.1 mm after 60 minutes.

The 19.8% (by mass) evaporated ANS did not spread as much initially as the fresh crude. The OP-40 performed slightly better than the TS6535 initially with the 19.8% evaporated ANS, but by the end of the experiments the TS6535 had maintained a 5.2 mm thickness while the OP-40 herded thickness had declined to 3.7 mm.

The 27.2 % evaporated ANS crude did not spread initially nearly as much as the less evaporated samples. The pour point of the 27.2% evaporated ANS would have been near 3°C. The water temperature at the start of the OP-40 experiment was −0.3° and at the start of the experiment with TS6535 was −1°C. Even at water temperatures slightly below the oil`s pour point, both herders contracted the slicks within one minute: OP-40 achieved a thickness of 6 mm and TS6535 achieved a thickness of 6.3 mm. Both herders maintained the initial thickness over the one-hour experiment. In general, both herders were more effective with ANS crude as evaporation increased.

The results of herding experiments with the ANS crude in pans with ≈30% slush ice (data not shown here – see Buist et al. 2017) simulated by adding a bag of crushed ice cubes to the water in the pan at the beginning of the experiment, showed that the presence of the slush ice restricted the initial spreading of the oil: initial equilibrium thicknesses were in the 1 to 2 mm range. The herders were effective on the fresh and 19.8% evaporated ANS crude in loose slush ice. The OP-40 herder was more effective on the fresh and 19.8% evaporated ANS crude than the TS6535. Neither herder had an appreciable effect on the 27.2% evaporated ANS in the slush ice conditions.

The results with the other three crudes were similar to these, with the following exceptions:

• The Endicott crude tests had to be performed at 20°C because of its higher pour point. Low water content Endicott emulsions (25%) were contracted by both herders, but less effectively than the water-free crude.

• The fresh and evaporated Terra Nova crude tests also were performed at 20°C due to its high pour point. The crude did not spread out initially as much as the ANS crude. The OP-40 herder was slightly more effective at herding the fresh and evaporated Terra Nova crude at first, but its effectiveness declined over the one-hour test so that by the end, both herders were containing about the same thickness of oil. As the Terra Nova evaporated, its pour point increased from 3°C to 21°C when it has lost 17.7% of its mass. Both herders could contract the unemulsified slicks of 17.7% evaporated crude at 20°C. The data from the experiments with emulsified Terra Nova show that, even though they spread less than the unemulsified crude initially, both herders could contract the emulsion slicks at 20°C. In other experiments at 10°C both herders managed to contract water-free and emulsified slicks of gelled Terra Nova crude at temperatures 8° to 11°C below their pour point.

• The initial equilibrium thicknesses for the unemulsified Grane crude at 0° C were in the 1 to 2 mm range. Both herders proved to be very effective with the Grane crude and emulsions, generally achieving herded thicknesses in the 5 to 7-mm range. The thickness of slicks herded with OP-40 tended to decline slowly over the one-hour test. The thickness of Grane slicks herded by the TS6535 tended to decline less, or not at all. Evaporated Grane crude formed fairly stable emulsions at 0°C and both herders contracted these slicks as well or better than the water-free slicks. The 50% water content slicks of the 10% evaporated Grane crude were not thickened as effectively as the less weathered slicks by either herder. Both herders were effective on the fresh and both evaporated Grane crude samples in slush ice. Initially, the OP-40 herder was somewhat more effective than the TS6535. At the end of the tests the OP-40 herded slicks of Grane crude in slush ice had re-spread slightly, declining in thickness to the levels maintained by the TS6535 herder.

Figure 3 shows the DFP results with the ANS crude. The ANS crude spread to an initial thickness of approximately 0.5 to 1.5 mm. The OP-40 herder appeared to work better on the fresh ANS than the TS6535, achieving an initial herded thickness of 3.5 mm vs. 3.1 mm; however, the OP-40 herded slick slowly spread out over the one-hour time span of the experiment, whereas the TS6535 maintained its initial herded thickness. This was the general trend with the evaporated ANS experiments as well. Both herders produced better results in the DFP test apparatus with the evaporated ANS crude than the fresh.

Fresh Terra Nova crude was also used to explore the effects of wave energy, pour point and emulsification on herder effectiveness in the DFP test. Even though the pour point of the fresh Terra Nova crude is 3°C, both herders were effective in contracting the slicks in the DFP apparatus at 0°C. In comparison, in the 1-m² pan tests the OP-40 was only slightly effective with Terra Nova at 0°C. This difference is likely due to the wave energy imparted to the slick and herder in the DFP apparatus which prevents a fully-developed precipitated wax matrix from forming gelled oil. This affords the herder less resistance in contracting the slicks. Both herders contracted emulsified slicks of fresh Terra Nova crude. The OP-40 was slightly more efficient than the TS6535.

Fresh and weathered Grane crude was also used in the DFP experiments. The initial equilibrium thicknesses for the Grane slicks were in the 1- to 2-mm range. Both herders caused significant contraction of all the Grane test slicks. Herded slick thicknesses were consistently in the 6 to 9 mm range, regardless of herder, evaporation or emulsification. These thicknesses were as good as, or slightly better than those achieved in the static 1-m² pan experiments.

The ANS and Grane crude samples tested in the 10-m² pool were selected based on the results of the 1-m² and DFP experiments. Some experiments involved ≈30% slush ice and some Grane crude experiments involved emulsions.

Figure 4 shows the results of the experiments with the two herders on fresh and evaporated ANS crude at 0°C. The initial equilibrium thickness of the ANS crude samples ranged from 0.2 to 0.5 mm, about half of the values measured in the 1-m² experiments (Figure 2). This is likely due to the 1-m² experiments involving 500 mL of crude on a 1 m² area of water while the 10-m² experiments used 1 L of crude on a 10 m² water surface.

The OP-40 worked better than the TS6535 with the fresh and two degrees of evaporation of the ANS crude. In general, the contracted slick thicknesses measured for both herders were quite similar in the 10-m² pool and in the 1-m² pans.

The initial equilibrium slick thicknesses in experiments with the Grane crude were again considerably lower in the 10 m² pool than in the 1 m² pan. Both herders were effective with this crude. The herded thicknesses in these tests were generally the same as in the 1-m² pans. Two emulsions produced with the 10% evaporated Grane crude were also tested in the 10-m² pool using the TS6535 herder. Both the 20% water and 50% water emulsions were effectively herded by the TS6535.

Some of the experiments with fresh ANS crude and fresh Grane crude in the 10-m² pool involved slush ice at a coverage of approximately 30%. The initial equilibrium thickness of both oils in slush ice was appreciably less in the 10-m² tests than in the 1-m² series. Both herders managed to contract the slicks in the presence of slush ice. The OP-40 herder worked better in the slush ice than the TS6535, and also appeared to work better in slush ice than on open water.

Table 1 presents the oil removal efficiencies measured in the six test burns. One burn with each fresh oil was conducted in a floating metal ring containing 400 mL of oil without herder, as a control. Theoretically, the 3.2 mm initial slick thickness (400 mL in a 40 cm diameter burn) should result in 68% removal efficiency. For each of the burns using herder, the slicks would expand slightly as the flames spread over the oil, then contract again as the flames began to extinguish.

The test burn with the fresh and evaporated ANS crude herded with OP-40 produced removal efficiencies of 59% for the fresh ANS, 46% for the 19.8% evaporated ANS and 56% for the 27.2% evaporated ANS. The lower efficiency measured for the 19.8% evaporated ANS is believed to have been due to experimental error. The burn with fresh ANS crude herded by TS6535 resulted in a removal efficiency of 49%.

The experiment with Grane crude herded with OP-40 produced a removal efficiency of 59% and the one involving fresh Grane herded by TS6535 resulted in a removal efficiency of 40%. There appears to be a trend of higher oil removal efficiency with OP-40 than with TS6535 in these experiments.

Four experiments in the wind/wave tank involved placing 500 mL of fresh crude in a monolayer of herder that had been previously applied to the water surface, then progressively increasing the steepness of the waves generated by the wave board. The oil was not ignited. The first experiment involved a slick of fresh ANS herded by TS6535. At the least energetic wave setting (10 cm wave height with a 2-s period) the herded slick would contract and expand as a wave passed under it. As wave periods were progressively decreased, the herded slick narrowed and divided into smaller slicklets. These slicklets were still constrained by the herder. When the wave period was shortened to 0.6 s, spilling breakers formed and quickly broke up the slicklets into small bits of oil surrounded by oil sheen. There were no noticeable slick behavior differences with fresh ANS crude herded by OP-40 in the second test. The third wave test involved fresh Grane crude herded by OP-40. For the lower wave energies, the Grane slick behaved in a manner similar to the ANS slick experiments. When the wave period was reduced to 0.6 s, the breaking waves quickly broke the slicklets up into small bits, but less sheening was observed than the ANS. The final experiment (Grane herded by TS6535) was similar.

Two experimental test pools were set up on leveled sand in the refrigerated Research Area at CRREL (Figure 5). The air temperature was maintained at −4°C. Each test pool was constructed using lumber to be approximately 5.3 m × 5.3 m × 15 cm. Each pool frame was laid on white cloth to permit good discrimination between oil, ice and water in the photos and video. Rinsed polyethylene film was used on top of the white sheeting and clamped to the sides to provide clean containment for each experiment. Cold salt water (35 ‰) was used to fill the test pools to a depth of 2.5 to 5 cm, enough to ensure the entire bottom of the pool was covered with at least 2 cm of water. The pool water surface tension was measured before each experiment to confirm cleanliness. A digital camera and a digital video camera mounted over the center of each pool were used to record the spread of the oil and subsequent herding for later analysis. Each experiment lasted for approximately one hour. The test oil was ANS crude, from CRREL supplies. This particular crude sample was also used in the SL Ross laboratory tests. Fresh, 17.5% evaporated and 29.4% mass evaporated crude were used at CRREL. Full procedure details may be found in the project report.

Figure 6 presents the results from the experiments on open water. Initial equilibrium thicknesses for the fresh, weathered and emulsified slicks were consistently 0.3 to 0.4 mm on the salt water in the pools. The OP-40 herder performed better than the TS6535, achieving thicknesses on par with those measured in the 10-m² experiments in the SL Ross laboratory (Figure 4). The thicker slicks seemingly obtained with OP-40 and the 20% water emulsion of the 29.4% evaporated ANS may be an artifact of the emulsion breaking shortly after being poured onto the water.

The performance of the TS6535 was particularly poor: appreciably less efficient than in the SL Ross laboratory experiment series (Figure 4). The experiments at CRREL used a different aliquot of the TS6535 than those at SL Ross, and it is possible that the sample sent to CRREL was deficient in some manner (possibly improperly mixed). After the first few tests, the TS6535 container was pre-warmed slightly and shaken vigorously before each 4 mL dose was withdrawn for a test. The duplicate experiment with the TS6535 and the fresh ANS was done after warming and shaking the herder container; whereas, the original experiment was conducted without the benefit of warming and mixing. There was an appreciable difference between the two tests, with the duplicate giving results much closer to those obtained with fresh ANS in the 1 m² pan and 10-m² pool at SL Ross. The 17.5% evaporated ANS was herded by the TS6535, but much more slowly and less effectively than in the SL Ross lab tests. Neither the 29.4% evaporated slick nor the 20% emulsified slick were appreciably herded by TS6535 over the one-hour experiments at CRREL. These tests were done after the procedure change was made to pre-warm and shake the metal can of TS6535. It is believed that the combination of the generally lesser performance of this batch of TS6535, the higher pour point of the most weathered ANS sample at CRREL (4.2°C vs. 3°C for the sample used at SL Ross) and rapid incremental evaporation from the sheening oil around the main body of the slick, explain the less effective herding of the most weathered ANS slicks. The rapid evaporation of the sheen may be the main reason. Although the air temperatures at CRREL were much colder than at SL Ross, the water temperatures were comparable at approximately 5°C. In reviewing the overhead video for these experiments it is clear that the sheen around the spreading oil is starting to gel even before it has ceased spreading, particularly for the 29.4% evaporated ANS.

In the first two tests with slush ice (ANS 17.5% evaporated with TS 6535 and ANS 17.5% evaporated with OP-40) the surface coverage of slush ice was much higher than intended, perhaps as high as 90 to 100%, when it spread after being deposited in the test pool. In both of these experiments the oil was prevented from spreading to its usual initial equilibrium thickness by the ice, and the ice prevented the herder from reaching the edge of the slicks. Adjustments were made to reduce the quantity of slush added to the test areas. In subsequent tests the OP-40 performed well in the 30% slush ice cover, with the fresh and 17.5% evaporated crude, albeit taking a little longer to thicken the slicks and achieving a lower one-hour herded thickness compared to the equivalent slicks in open water. The herded 29.4% evaporated slick was consistently thinner than the equivalent in open water, but did achieve an ignitable thickness after 30 minutes. The same was true for the emulsified 29.4% evaporated crude. The TS6535 was not nearly as effective as the OP-40 in 30% slush ice: it barely contracted the fresh oil to an ignitable thickness after one hour and had no discernible effect on the 17.5%, 29.4% or emulsion slicks.

• OP-40 was generally better than TS6535 as an oil slick herder.

• Both herders could contract thicker oil slicks at water temperatures up to approximately 10° C below their pour point. Since the oil gelling process reduces evaporation rates of oil on water, this may extend the window of opportunity for herding and burning in cold waters. At temperatures more than 10° C below the oil`s pour point, the herder will not contract the slick.

• Gentle wave action appears to assist when herders are acting on a slick at temperatures up to 10 degrees below its pour point.

• High concentrations of slush ice restrict oil spreading and prevent herders from reaching the edge of the slick.

• Herders can thicken oil slicks in low slush ice concentrations (basically, if the oil can spread, the best herders can contact the slick edge and cause it to contract) and maintain ignitable slick thicknesses for an hour.

• High pour point oils in slush ice are not contracted by herders as well as on water.

• Herders contract low water content emulsions (e.g., 20%). They do not generally work well on higher water content emulsions (50%), but these are rarely ignitable.

• The effectiveness of herders on oil generally increases with increasing evaporation and decreases with increasing water content of the oil slick. Herders are known to be less effective on lighter, volatile oils and the loss of these components from the crude may explain this.

This paper was prepared under contract for the International Association of Oil & Gas Producers (IOGP) by S.L. Ross Environmental Research Ltd as prime contractor. Publication of this paper does not necessarily imply that the contents reflect the views and policies of IOGP, nor are there any implied IOGP endorsements of studies performed and results presented by SL Ross or by the subcontracting partners DCE and CRREL.